What causes CLOUD_PROVIDER_LAUNCH_FAILURE?

The requested VM instance type is not available in the selected availability zone or region. IAM role, service principal, or managed identity lacks permissions to launch EC2/VMs. Cloud account quota for the instance type or vCPU count has been exceeded. The AMI, VM image, or custom image ID referenced in the cluster config is invalid or not accessible. An AWS/Azure/GCP service outage in the target region

How do I fix CLOUD_PROVIDER_LAUNCH_FAILURE?

Check the cloud provider console for errors on the VM launch request. Verify the IAM role or service principal has the required EC2/VM launch permissions. Check cloud account quotas for vCPUs and request an increase if needed. Try a different availability zone or use instance type fallback in the cluster configuration. Check the cloud provider status page for active incidents in your region

High severitycluster

Power BI Refresh Error:
CLOUD_PROVIDER_LAUNCH_FAILURE

What does this error mean?

The cloud provider (AWS, Azure, or GCP) failed to launch the VM instances requested for the Databricks cluster. The error is at the cloud infrastructure level, not in Databricks itself.

Common causes

1The requested VM instance type is not available in the selected availability zone or region
2IAM role, service principal, or managed identity lacks permissions to launch EC2/VMs
3Cloud account quota for the instance type or vCPU count has been exceeded
4The AMI, VM image, or custom image ID referenced in the cluster config is invalid or not accessible
5An AWS/Azure/GCP service outage in the target region

How to fix it

1Check the cloud provider console for errors on the VM launch request
2Verify the IAM role or service principal has the required EC2/VM launch permissions
3Check cloud account quotas for vCPUs and request an increase if needed
4Try a different availability zone or use instance type fallback in the cluster configuration
5Check the cloud provider status page for active incidents in your region

Frequently asked questions

How do I find the specific AWS or Azure error?

Check the cluster event log in Databricks — it usually includes the raw error message from the cloud provider, which gives the specific reason for the launch failure.

Can I configure fallback instance types?

Yes — Databricks supports instance pool fallback and spot-to-on-demand fallback configuration to reduce the impact of capacity shortages.